Mapillary Vistas Dataset

该数据集名为Mapillary，包含了25,000张图像，被用作测试模型的靶向数据集。此外，它还是一个用于评估语义分割的现实世界数据集。该数据集的规模为25,000张图像，所涉及的任务是语义分割。

This dataset, named Mapillary, contains 25,000 images and serves as a targeted dataset for model testing. Moreover, it is a real-world dataset dedicated to semantic segmentation evaluation. With a total of 25,000 images, the core task involved in this dataset is semantic segmentation.

搜集汇总

背景与挑战

背景概述

Mapillary Vistas Dataset是一个全球街景图像数据集，包含25,000张高分辨率图像，覆盖124个语义类别和100个实例特定标注，适用于街景理解研究。数据集提供多样化的拍摄条件和精细标注，遵循CC BY-NC-SA非商业许可协议。